1. Field of the Invention
The present invention relates to a recording apparatus, and more particularly to an apparatus for recording image signals encoded through interframe prediction.
2. Related Background Art
Digital data processing is vigorously studied nowadays. Various methods for standardization, particularly of high efficiency coding for image data compression have been proposed and discussed. Of these, a general coding scheme MPEG 2 (Moving Picture Coding Expert Group) has drawn attention which is adopted to ATV (Advanced Television) of next generation TV broadcast of U.S.A.
MPEG 2 is a motion compensation type prediction coding method of compressing information amount and coding data by utilizing correlation between image frames. In FIG. 1, arrows indicate directions of coding predication. FIG. 2 is a diagram illustrating coding by MPEG 2, data order on media, and image data order in decoding.
In MPEG 2, a GOP (Group of Pictures) is constituted of a predetermined number of frames. Each GOP includes at least one frame of intraframe encoded images.
An intraframe encoded image (hereinafter called an I image) is an image encoded by using image data only in one frame, through DCT and quantization. Image data at every Nth frame from the I image is encoded through interframe prediction coding (hereinafter called a P image). Image data of each frame between I and P images and between P images is encoded through bidirectional prediction coding by using image data of backward and forward frames (hereinafter called a B image).
As shown in FIGS. 1 and 2, first I images are formed. As described above, an I image is encoded by using image data of only one frame, and prediction using data of other frames is not performed. Next, P images are formed, and B images are formed after I or P image. The I, P, and B images are transmitted in this order.
In MPEG 2, a data stream of encoded image/voice data or other data is called an elementary stream. As a structure for transmitting an elementary stream, a PES (Packetized Elementary Stream) packet has been defined. This structure has a PES payload (data field) following a PES header. In MPEG 2, a set of elementary streams having the common time axis is called a program.
Two formats have been defined for multiplex in MPEG 2. One is a transport stream and the other is a program stream.
Both definitions of the transport stream and the program stream contain necessary and sufficient syntax for synchronizationronization when decoding and reproducing images and voices. The program stream is a single data stream obtained by connecting one or more PES packets having the common time axis. The transport stream is a single data stream obtained by connecting one or more programs having one or more time axes. The above mentioned ATV uses the transport stream.
In the transport stream, data of images, voices, and the like is transmitted by dividing it into a transmission unit of a fixed length of 188 bytes called a transport packet.
Information such as various identifiers called a PCR (Program Clock Reference) and a PSI (Program Specific Information) used for synchronization is inserted where necessary into the transport stream. This information is detected to correctly decode the encoded data.
PSI contains information for discriminating a PID (Packet ID) or the like, called a PAT (Program Association Table) or a PMT (Program Map Table). This information is used for detecting and decoding a packet containing a target program or data.
As described earlier, since the I image is encoded by using image data of only one frame, the encoded data can be decoded by using only this encoded data. On the other hand, since the P and B images are encoded by using image data of other frames, the encoded data cannot be decoded by using only the encoded data.
The data length of each of I, P, and B image data is variable. Therefore, if image data encoded by MPEG 2 is recorded in a recording medium such as a magnetic tape, the location of the I image on the recording medium cannot be identified.
For example, if image data encoded by MPEG 2 is recorded in a magnetic tape by using an apparatus such as a digital VTR, the original image data can be correctly reproduced if it is reproduced at the same speed as recording, because the image data is reproduced in the order of recording, i.e., in the order of encoding.
However, in specific reproduction such as high speed search, the head transversely traces a tape so that each encoded image I, P, and B is not reproduced in the order of encoding. Furthermore, since each head traces only a fraction of each track, the location of the I image on the tape cannot be identified as described above and the I image is not necessarily reproduced correctly. Therefore, specific reproduction is associated with a problem that image data cannot be reproduced correctly.